ABSTRACT Clinical decision making will increasingly depend on validated, high-quality biomarkers that can be used to guide cancer surveillance and tailor appropriate treatment. Identifying colorectal cancer (CRC) markers is particularly critical, as CRC is common in both men and women (8% of all incident USA cancers) and is frequently lethal (USA 5-year survival rate: 65%). To date, clinically useful biomarkers predictive of recurrence or survival for CRC patients are limited. Treatment decisions are based largely on clinical and pathologic parameters, with little else to guide risk and treatment stratification of patients. This study will focus on metabolite biomarker discovery and validation utilizing 1,840 patients (stage I-III) with 6,004 repeat blood samples from the existing ColoCare Study, a multi-center prospective cohort of newly diagnosed CRC patients, including detailed demographic, clinical, epidemiologic, and follow-up data, for which blood samples are collected at multiple standardized time points, using identical protocols across study sites. This diverse population ensures broad generalizability and clinical applicability of identified biomarkers. CRC is known to affect metabolism, and thus it is anticipated that markers of altered metabolism should yield useful diagnostic information. Our discovery of metabolic biomarkers will yield novel, distinct findings, and will also be synergistic with ongoing, separate analyses of proteomic, glycomic, and autoantibody biomarkers in the ColoCare Study patient population. This project?s specific aims are: To use state-of-the art, well-validated metabolomic platforms to discover and verify novel biomarkers: 1. Predictive of recurrence among CRC patients: Using samples collected at diagnosis and follow-up, we will identify metabolites predictive of risk of recurrence in stage I/II and stage III patients. 2. Capable of early detection of CRC recurrence: Using serial samples collected at regular post-surgical intervals (6, 12, and 24 months), we will identify biomarkers useful for disease monitoring for recurrence. Metabolite biomarkers will include >2,500 lipids and aqueous metabolites (MW<1,000; distinct from proteins). We will evaluate the performance of identified markers separately for men and women and perform analyses to understand factors that affect their performance. Our long-term goal is to develop clinical-grade biomarker assays that have a significant impact on reducing morbidity and mortality associated with CRC through guidance of treatment/follow-up decision making and characterization of risk of recurrence. The proposed metabolomics platforms are state-of-the art, have yielded potentially useful biomarkers in the past, and have not yet been used in the context of CRC prognosis. The study uses a rigorous multi-step design and is expected to yield clinically robust markers ready for rapid translation.